We observe glitches in the arrival times of X-ray pulses from GRO J174
4-28 associated with the large X-ray bursts observed from this source.
In early observations, the pulse arrival time becomes delayed by 28 m
s (6% of the 467 ms pulsar period) following bursts while the shape of
the nearly sinusoidal light curve is essentially unchanged. This arri
val-time lag decays exponentially over the next similar to 1000 s. In
later observations, the observed phase lags are smaller, and eventuall
y, some bursts appear to be accompanied by a small phase advance. Whil
e explanations of these observations that rely upon changes in the pos
tburst accretion flow are natural, our observations present some probl
ems for these explanations. It may be that this behavior can be unders
tood by analogy with glitches in spinning-down radio pulsars, though t
his explanation has its own problems. Explained as a glitch, the phase
shift is the result of coupling of the neutron star's crust to the cr
ustal superfluid, which is spun up more slowly than the rest of the st
ar, and the recovery is due to coupling between the crust and the core
of the neutron star.